JP4367111B2 - Semiconductor dynamic quantity sensor - Google Patents

Description

  The present invention relates to a semiconductor dynamic quantity sensor, and more particularly to a sensor element formed by etching a semiconductor crystal having an active layer superimposed on a support substrate via an insulating layer, and an electrode pad connected to the outside. The present invention relates to a semiconductor dynamic quantity sensor including wiring to be connected.

  2. Description of the Related Art Conventionally, a semiconductor dynamic quantity sensor including a sensor element including a movable part that can be displaced is known (see, for example, Patent Document 1). In such a semiconductor dynamic quantity sensor, an electrode is provided for electrically detecting the displacement amount of the movable part or forcibly vibrating the movable part. These electrodes are connected to an electrode pad connected to the outside by a wire or the like, and output a sensor signal from the electrode pad to an external electric circuit or send a signal supplied from the external electric circuit via the electrode pad. input.

In the configuration described above, wiring is interposed between the electrode as the sensor element and the electrode pad. Like the electrodes, this wiring is formed by etching a semiconductor crystal having an active layer superimposed on a support substrate via an insulating layer. In the above conventional semiconductor dynamic quantity sensor, this electrode and electrode The wiring connecting the pad is fixed to the support substrate with the insulating layer remaining below the wiring.
JP 2000-193460 A

  By the way, in order for the insulating layer below the wiring to remain even if the etching process is performed as described above, it is necessary to relatively increase the thickness of the wiring. However, when the wiring becomes thicker, the exclusive space for the wiring is increased accordingly, and it is difficult to reduce the size of the sensor itself. On the other hand, if the wiring thickness is reduced to such an extent that the insulating layer under the wiring is completely removed by the etching process, the space occupied by the wiring can be reduced and the sensor can be downsized. In some cases, the insulating layer no longer exists below the wiring, so that the wiring floats from the support substrate, and a situation may occur where the wiring contacts the support substrate due to the deformation or displacement thereof. When the wiring is in direct contact with the support substrate, it is impossible to properly extract the sensor signal or vibrate the movable part.

  The present invention has been made in view of the above-described points, and an object of the present invention is to provide a semiconductor dynamic quantity sensor capable of suppressing the contact of the wiring with the support substrate while reducing the exclusive space of the wiring as much as possible. And

The above object is achieved by etching a semiconductor crystal in which an active layer is stacked on a support substrate via an insulating layer, and includes a semiconductor dynamic quantity including wiring for connecting a sensor element and an electrode pad connected to the outside. In the sensor, a widened portion is provided in the middle of the wiring, which has a width larger than that of the main body of the wiring, and is fixed to the support substrate via the insulating layer, and the widened portion is disposed on the wiring. And the predetermined length is set to such a length that the sticking force of the wiring to the support substrate side is smaller than the spring reaction force in the sticking direction of the wiring. Is achieved.

  In the first aspect of the present invention, a part of the wiring connecting the sensor element and the electrode pad is provided with a widened portion having a width larger than that of other portions. In such a configuration, the insulating layer under the widened portion remains without being removed by etching, even if the wiring portion is formed thin so that the insulating layer is lifted from the support substrate by etching away. Is possible. In this case, the wiring is fixed to the support substrate via the insulating layer at the widened portion, and is floated from the support substrate at another portion. Therefore, by reducing the thickness of the wiring, the exclusive space for the wiring can be reduced, and the presence of the widened portion can suppress the contact of the wiring with the support substrate.

In the present invention, since the widened portion is provided for each predetermined length on the wiring, it is possible to reliably prevent contact of the main body portion having a relatively small width with respect to the support substrate.
In this case, in the above-described semiconductor dynamic quantity sensor, the main body portion is floated from the support substrate by the etching of the insulating layer below the main body portion , and the widened portion is etched by the insulating layer below the main body portion. In other words, it may be fixed to the support substrate via the insulating layer.

In this case, in the above-described semiconductor dynamic quantity sensor, the predetermined length is set to a length in which a sticking force of the wiring to the support substrate side is smaller than a spring reaction force in the sticking direction of the wiring. do it.
In the semiconductor dynamic quantity sensor described above, the predetermined length for the non-linear portion of the wiring may be set to be shorter than the predetermined length for the linear portion of the wiring.
In the semiconductor dynamic quantity sensor described above, the main body may have a thickness that is approximately the same as the thickness of the movable electrode of the sensor element.
Furthermore, in the semiconductor dynamic quantity sensor described above, the widened portion may be formed in a square shape.

According to the present invention, it is possible to reduce the exclusive space of the wiring by thinning the wiring so that it floats from the support substrate, and to suppress the contact of the wiring with the support substrate due to the presence of the widened portion.

Moreover, according to this invention, the contact to the support substrate of the main-body part with comparatively small width | variety among wiring can be prevented reliably.

  FIG. 1 is an overall plan view of a semiconductor dynamic quantity sensor 10 according to an embodiment of the present invention. The semiconductor dynamic quantity sensor 10 of this embodiment is a semiconductor sensor that is mounted on, for example, a vehicle or the like and detects an angular velocity generated around the center of gravity axis of the vehicle.

  In this embodiment, the semiconductor dynamic quantity sensor 10 includes a sensor element 12. The sensor element 12 includes a substantially rectangular support substrate made of silicon, an insulating layer (sacrificial layer) made of a silicon oxide film having a predetermined thickness formed on the support substrate, and silicon formed on the insulating layer. It is formed by performing etching by microfabrication on the surface of the silicon semiconductor crystal laminated with the active layer. The sensor element 12 includes a vibrator 14 made of silicon having a predetermined mass. The vibrator 14 is in a state of floating from the support substrate by removing the insulating layer below it by etching (sacrificial layer etching), and is connected to an anchor 16 fixed to the support substrate via beams 18 and 20. And is supported by a support substrate. The beam 18 is a beam that allows displacement of the vibrator 14 relative to the anchor 16 in the X-axis direction in the figure. The beam 20 is a beam that allows displacement of the vibrator 14 relative to the anchor 16 in the Y-axis direction in the figure. Accordingly, the vibrator 14 is configured to vibrate in both the X-axis direction and the Y-axis direction with respect to the support substrate.

  The sensor element 12 also has a drive electrode portion 22, a drive monitor electrode portion 24, a detection electrode portion 26, a detuning adjustment electrode portion 28, and a detection direction drive electrode, each formed in a comb shape having a plurality of electrode fingers. The unit 30 is provided. Each of these electrode portions 22 to 30 has a movable electrode provided integrally with the vibrator 14 and a fixed electrode fixed on the support substrate. In each of the electrode portions 22 to 30, the movable electrode and the fixed electrode are arranged so as to face each other with a predetermined gap in a normal state.

  The drive electrode unit 22 is configured to drive the vibrator 14 to vibrate in the X-axis direction with respect to the support substrate by applying an electrostatic attractive force between the fixed electrode and the movable electrode by applying an excitation voltage to the fixed electrode. It constitutes an electrode. The drive monitor electrode unit 24 drives the vibrator 14 in the X-axis direction relative to the support substrate by detecting a change in capacitance between the fixed electrode and the movable electrode accompanying the displacement of the vibrator 14 in the X-axis direction. The electrode for monitoring is comprised. The detection electrode unit 26 detects a change in electrostatic capacitance between the fixed electrode and the movable electrode due to the displacement of the vibrator 14 in the Y-axis direction, thereby detecting the vibration displacement of the vibrator 14 in the Y-axis direction with respect to the support substrate. An electrode for detection is configured. Hereinafter, the X-axis direction will be referred to as the drive direction of the vibrator 14 and the Y-axis direction will be referred to as the detection direction of the vibrator 14 as appropriate.

  In addition, the detuning adjustment electrode unit 28 applies an electrostatic attractive force between the fixed electrode and the movable electrode by applying a DC bias voltage to the fixed electrode, thereby detecting the resonance frequency of the detected vibration in the detection direction of the vibrator 14. To change the gain of the detected vibration (effective detected vibration Q value). The detuning adjustment refers to adjusting a difference between the resonance frequency of the vibration detected by the vibrator 14 and the resonance frequency in the driving direction. The detection direction drive electrode unit 30 causes the vibrator 14 to move in the Y-axis direction (detection direction) with respect to the support substrate by applying an electrostatic attractive force between the fixed electrode and the movable electrode by applying an excitation voltage to the fixed electrode. The electrode for making it drive is comprised.

  A base part 34 formed on the insulating layer is connected to the fixed electrodes of the electrode parts 22 to 30 through wirings 32. On the upper surface of each base portion 34, a rectangular electrode pad 36 for taking out a potential to the outside made of a conductive metal such as aluminum or taking in a potential from the outside is disposed. A connection conducting wire (wire) made of aluminum or the like is joined to the upper surface of the electrode pad 36.

  In this embodiment, a rectangular glass lid is fixed to the upper surface of the sensor element 12 so as to cover the vibrator 14, the beams 18 and 20, the various electrode portions 22 to 30, the wiring 32, and the electrode pad 36. ing. In such a structure, the vibrator 14 is configured to be able to vibrate in a space sealed by the support substrate, the insulating layer, the active layer, and the glass lid.

  Next, the manufacturing method of the semiconductor dynamic quantity sensor 10 of the present embodiment will be described in the order of steps.

  First, an SOI (Silicon On) formed by laminating a support substrate made of silicon, an insulating layer made of a silicon oxide film having a thickness of several μm, and an active layer made of silicon having a thickness of several tens of μm on the upper surface thereof. Insulator) A silicon oxide film is formed on the entire upper surface of the wafer (first step). Next, patterning of the silicon oxide film on the SOI wafer is performed, and a portion corresponding to the vibrator 14, the electrode portions 22 to 30, the base portion 34 wiring 32, and the beams 18 and 20 and a slight width are added to this portion. Are formed (second step). Then, the active layer is etched by RIE (reactive ion etching) or the like to form a through hole in a portion that will later become the vibrator 14, and the base portion 34, the fixed electrodes of the electrode portions 22 to 30 on the insulating layer, Then, a frame is formed, and portions corresponding to the vibrator 14, the movable electrodes of the electrode portions 22 to 30, the wiring 32, and the beams 18 and 20 are left (third process).

  Next, the wafer is immersed in a bath filled with a hydrofluoric acid aqueous solution (etching solution) that dissolves silicon oxide but does not dissolve silicon, and the vibrator 14, the movable electrodes of the electrode portions 22 to 30, the wiring 32, and The insulating layer sandwiched between the beams 18 and 20 and the support substrate is removed by isotropic etching to form the vibrator 14, the movable electrodes of the electrode portions 22 to 30, the wiring 32, and the beams 18 and 20 (fourth). Process). Then, an aluminum film is formed on the portion corresponding to the electrode pad 36 by a sputtering method or the like to form the electrode pad 36 (fifth step). Next, a connecting wire (wire) W made of aluminum is bonded to the upper surface of the electrode pad 36 by an ultrasonic wire bonding method or the like, and these connecting wires are connected to terminals of an electric circuit (not shown) (seventh step). ). Finally, the above-described glass lid (not shown) is fixed to the upper surface of the frame body by anodic bonding or the like in the vacuum (eighth step).

  Next, the operation of the semiconductor dynamic quantity sensor 10 of this embodiment as a sensor will be described. In this embodiment, the electric circuit transmits driving signals of opposite phases to the fixed electrode of the driving electrode portion 22 via the electrode pad 36 in order to vibrate the vibrator 14 at its natural frequency and with a constant amplitude in the X-axis direction. Supply. Further, in order to detect the vibration of the vibrator 14 in the Y-axis direction, detection signals having opposite phases to each other are supplied to the fixed electrode of the detection direction drive electrode 30 via the electrode pad 36.

  When the driving signal as described above is supplied to the fixed electrode of the drive electrode unit 22, an electrostatic attractive force is generated between the fixed electrode and the movable electrode of the drive electrode unit 22, so that the vibrator 14 has a natural frequency. Vibrates with a constant amplitude in the X-axis direction. When an angular velocity about the Z-axis (the direction penetrating the paper surface in FIG. 1) perpendicular to the X-axis and also perpendicular to the Y-axis is generated in the excited state of the vibrator 14, the vibrator 14 has a vibration speed Vx in the X-axis direction. , Mass m, and Coriolis force Fc = 2m · Vx · Ωz in the Y-axis direction (detection direction) according to the angular velocity Ωz acts. When such a Coriolis force is applied, the vibrator 14 is displaced in the Y-axis direction, and vibrates with an amplitude proportional to the magnitude of the angular velocity Ωz and a period according to the frequency of the drive vibration in the X-axis direction.

  When the vibrator 14 vibrates in the Y-axis direction, the gap between the movable electrode and the fixed electrode of the detection electrode unit 26 changes by the magnitude of the Coriolis force. Further, since the movable electrodes of the two detection electrode portions 26 formed integrally with the vibrator 14 vibrate in the same direction in the Y-axis direction, the capacitance between one movable electrode and the fixed electrode, and the other The electrostatic capacitance between the movable electrode and the fixed electrode changes in opposite directions. These signals representing changes in capacitance are input to the electric circuit via the electrode pads 36 of the two detection electrode portions 26. The electric circuit detects the angular velocity around the Z axis based on the input electrostatic capacitance signal.

  By the way, in the semiconductor dynamic quantity sensor 10 of the present embodiment, the wiring 32 connecting the fixed electrodes of the sensor element 12 and the electrode pads 36 for connecting the electrodes to an external electric circuit are interposed. However, like the electrodes, these wirings 32 are formed by etching a silicon semiconductor crystal in which an active layer is stacked on a support substrate via an insulating layer.

  However, if the insulating layer below the wiring 32 is left even if the etching process is performed, and the entire wiring 32 is fixed to the support substrate through the insulating layer, the thickness of the wiring 32 is reduced by the etching process. Since it is necessary to make the thickness of the movable electrode larger than the thickness of the movable electrode from which the lower insulating layer is removed, the space occupied by the wiring 32 is increased correspondingly, making it difficult to reduce the size of the sensor 10 itself. On the other hand, in order to avoid the above inconvenience, if the thickness of the wiring 32 is made as thin as the movable electrode so that all the insulating layers below the wiring are removed by the etching process, the exclusive space of the wiring 32 is reduced. However, since the insulating layer under the wiring is removed by the etching process, the wiring 32 floats from the support substrate, and the wiring 32 and the support substrate are brought into contact with each other due to the bending deformation and displacement. As a result, it becomes impossible to appropriately extract the sensor signal from the electrode and supply the drive signal to the electrode.

  Therefore, the semiconductor dynamic quantity sensor 10 of the present embodiment has a structure that should avoid the above-described disadvantages. Hereafter, the characteristic part of a present Example is demonstrated with reference to FIG. 2 thru | or FIG.

  FIG. 2 is a plan view of the main part of the semiconductor dynamic quantity sensor 10 of the present embodiment. FIG. 3 shows a cross-sectional view of the semiconductor dynamic quantity sensor 10 shown in FIG. FIG. 4 is a diagram in which the method for manufacturing the wiring widening portion, which is a characteristic structure of the present embodiment, is arranged in the order of steps. 3A shows a cross-sectional view taken along line III-III, and FIG. 3B shows a cross-sectional view taken along line IV-IV.

  As shown in FIG. 2, in the semiconductor dynamic quantity sensor 10 of the present embodiment, the wiring 32 connecting each fixed electrode of the sensor element 12 and the electrode pad 36 is wider than the wiring body 32a and the wiring body 32a ( And a wide wiring widening portion (anchor) 32b having a rectangular shape. That is, in the middle of the wiring 32, a wiring widening portion 32b having a width larger than that of the wiring main body 32a is provided. The wiring main body 32a has the same thickness as that of each movable electrode of the sensor element 12, and the insulating layer 42 below the wiring main body 32a is removed by isotropic etching, as shown in FIG. In this way, the substrate floats from the support substrate 40. On the other hand, the wiring widened portion (anchor) 32b has a thickness larger than the thickness of each movable electrode of the sensor element 12, and a part of the insulating layer 42 below remains even by isotropic etching. As a result, as shown in FIG. 3B, the substrate is fixed to the support substrate 40 with the insulating layer 42 interposed therebetween.

  That is, in the sensor element 12, specifically, the wiring main body 32a portion and the wiring widening portion 32b portion are distinguished from each other in the patterning on the semiconductor crystal in the manufacturing stage of the wiring 32. If the wiring widening portion 32b is relatively thick (FIG. 4A), the width (thickness) of the wiring widening portion 32b is larger than that of the wiring main body 32a. Then, when isotropic etching with a solution is performed, the wiring body 32a is lifted from the support substrate 40 by the removal of the insulating layer 42 below the wiring body 32a. By partially remaining the insulating layer 42 below the widened portion 32b, the wiring widened portion 32b is fixed to the support substrate 40 through the insulating layer 42 (FIG. 4C).

  In such a structure of the semiconductor dynamic quantity sensor 10, the thickness of the wiring body 32 a can be reduced to such an extent that the wiring body 32 a floats from the support substrate 40. The space occupied by the wiring 32 can be reduced as compared with the case of being fixed. Further, when the wiring main body 32a floats from the support substrate 40 in this way, there is a possibility that the wiring main body 32a may come into contact with the support substrate 40. Since the portion 32 b is fixed to the support substrate 40 via the insulating layer 42, the contact of the wiring 32 including the wiring body 32 a floating from the support substrate 40 to the support substrate 40 can be suppressed.

  As the length (wiring length) L of the wiring 32 between the wiring widened portions 32b increases, the wiring main body 32a that floats from the supporting substrate is more likely to bend toward the supporting substrate, and the wiring main body 32a contacts the supporting substrate. It becomes easy to do. In this respect, in order to reliably prevent the wiring 32 from contacting the support substrate 40, it is necessary to appropriately set the length (wiring length) L of the wiring 32 between the wiring widened portions 32b. It is necessary to grasp whether the wiring widening portion 32b is provided for each L. Hereinafter, the optimum value of the wiring length L between the wiring widened portions 32 will be described. However, the optimum value varies depending on the material of the active layer constituting the semiconductor crystal, the shape of the wiring 32, and the like. Hereinafter, a case where the active layer is made of silicon will be described.

  FIG. 5 is a diagram showing the wiring 32 and the dimension below it. FIG. 6 is a diagram showing an analysis result of the wiring length L between the wiring widened portions 32b. In order for the wiring main body 32a of the wiring 32 not to contact the support substrate 40, a sticking force (sticking force) Fs acting as a force when the wiring main body 32a tries to deform / displace toward the support substrate 40 side is reduced. What is necessary is just to be smaller than the spring reaction force Fk in the sticking direction of the wiring 32 which shows the rigidity of the wiring main body 32a between 32b. That is, Fs / Fk <1 should be satisfied.

  The sticking force Fs described above includes the pressure PL acting on the surface of the liquid accumulated between the wiring body 32a, which is an active layer after the isotropic etching, and the support substrate 40, and the area A of the wiring body 32a between the wiring widening portions 32b. (= L × W; L is the wiring length described above, and W is the width (thickness) of the wiring body 32a). The pressure PL described above depends on the surface tension γ of the liquid, the contact angle θ between the wiring main body 32a and the liquid, the opposing distance between the lower surface of the active layer and the support substrate 40, that is, the thickness d of the insulating layer 42. It can be expressed as equation (1).

PL = 2 × γ × cos θ / d (1)
The spring reaction force Fk described above can be expressed by the product of the spring constant k of the wiring 32 and the thickness d of the insulating layer 42 between the wiring widened portions 32b. This spring constant k can be expressed by the following equation (2) by the width W, the vertical thickness (active layer thickness) T, and the wiring length L of the wiring body 32a.

k = W × T ³ / L ³ (2)
Therefore, when the relationship expressed by the following expression (3) is satisfied, the contact of the wiring body 32a with the support substrate 40 can be reliably prevented.

Here, γ = 20 × 10 ⁻³ [N / m] and cos θ = 0.5 are assumed, respectively, and the thickness (opposite distance) d of the insulating layer 42 is 4 [μm]. When the wiring length L is changed to 500 [μm], 1500 [μm], 2000 [μm], and 2500 [μm] under the condition that T is fixed to 40 [μm], the spring constant k is set to the wiring length. Since it is inversely proportional to the cube of L, the sticking force Fs and the spring reaction force Fk change as shown in FIG. In this analysis, the wiring length L satisfying the relationship of Fs / Fk <1 between the sticking force Fs and the spring reaction force Fk is 500 [μm] and 1500 [μm].

  Therefore, when the semiconductor dynamic quantity sensor 10 of the present embodiment has such a structure, if the wiring length L between the wiring widening portions 32b is set to about 1500 μm at the maximum, If the wiring widening portion 32b is installed at a maximum of every 1500 μm, it is possible to surely prevent the wiring main body 32a from being lifted from the supporting substrate 40 from contacting the supporting substrate 40. For the straight line portion of the wiring 32, it is effective to set the wiring length L between the wiring widened portions 32b according to the above-described analysis results. However, for the non-linear portions such as the L-shaped portion and the curved portion of the wiring 32, It is appropriate to set the wiring length L shorter than that for the straight line portion.

  Thus, according to the semiconductor dynamic quantity sensor 10 of the present embodiment, the wiring 32 is constituted by the relatively thin wiring body 32a and the relatively thick wiring widening portion 32b, and the wiring length L between the wiring widening portions 32b is optimal. By setting the value, it is possible to reliably prevent the wiring 32 from contacting the support substrate 40 while reducing the exclusive space of the wiring 32 as much as possible. For this reason, in the present embodiment, the function of the semiconductor dynamic quantity sensor 10 can be appropriately ensured while downsizing.

  In the above-described embodiment, the wiring widening portion 32b corresponds to the “widening portion” described in the claims, and the wiring main body 32a corresponds to the “other portion” described in the claims. .

  In the above embodiment, a semiconductor sensor for detecting an angular velocity around the Z axis is used as the semiconductor dynamic quantity sensor 10, but the present invention is not limited to this, and acceleration and deceleration are not limited thereto. It is also possible to apply to a semiconductor sensor that detects a mechanical quantity such as.

1 is an overall plan view of a semiconductor dynamic quantity sensor according to an embodiment of the present invention. It is a principal part top view of the semiconductor dynamic quantity sensor of a present Example. It is sectional drawing at the time of cut | disconnecting the semiconductor dynamic quantity sensor shown in FIG. It is the figure which arranged the manufacturing method of the wiring widening part which the semiconductor dynamic quantity sensor of a present Example has in order of the process. It is the figure showing the wiring which a semiconductor dynamic quantity sensor has, and its periphery dimension. It is a figure showing the analysis result about the distance between the wiring widening parts in a semiconductor dynamic quantity sensor.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Semiconductor mechanical quantity sensor 12 Sensor element 32 Wiring 32a Wiring main body 32b Wiring widening part 36 Electrode pad 40 Support board 42 Insulating layer

Claims (5)

A semiconductor dynamic quantity sensor formed by etching a semiconductor crystal having an active layer superimposed on a support substrate via an insulating layer, and comprising a wiring for connecting a sensor element and an electrode pad connected to the outside,
In the middle of the wiring, there is provided a widened portion that has a larger width than the main body of the wiring and is fixed to the support substrate via the insulating layer,
The widened portion is installed on the wiring every predetermined length , and
The semiconductor dynamic quantity sensor according to claim 1, wherein the predetermined length is set to a length in which a sticking force of the wiring to the support substrate side is smaller than a spring reaction force in a sticking direction of the wiring .   2. The semiconductor dynamic quantity sensor according to claim 1, wherein the predetermined length for the non-linear portion of the wiring is set to be shorter than the predetermined length for the linear portion of the wiring. The main body portion floats from the support substrate when the insulating layer below the main body portion is removed by etching, and
2. The semiconductor dynamic quantity sensor according to claim 1, wherein the widened portion is fixed to the support substrate via the insulating layer when the insulating layer below the remaining portion is left by etching.   The semiconductor dynamic quantity sensor according to claim 1, wherein the main body portion has a thickness approximately equal to a thickness of the movable electrode of the sensor element.   2. The semiconductor dynamic quantity sensor according to claim 1, wherein the widened portion is formed in a square shape.

Images